2010 제7회 FANEA동북아농정 연구포럼

국제심포지엄The 7th FANEA International Symposium

• 일시 : 2010. 6. 10(목), 09:00~18:00• 장소 : 제주국제컨벤션센터• 주최 : 한국농촌경제연구원(KREI), 중국농업과학원 농업경제발전연구소(IAED/CAAS), 일본 농림수산성 농림수산정책연구소(PRIMAFF)

The 7th FANEA Symposium

1. Date: 2010 June 10th

2. Place: Jeju International Convention Center

3. Time Schedule

9:00-9:20 Registration

9:20-9:40 Opening Ceremony

❍ Opening Remark: Dr. SeIk OH, President of KREI

❍ Welcoming Remark: The Governor of Jeju Special Self-governing Province

❍ Congratulatory Remark: The president of Jeju Development Institute

9:40-9:50 Break Time

9:50-11:10 Agenda 1 presentation: Agricultural Outlook using Econometric Model

Moderator: Dong Yang WANG, Vice DG from China

✔ Outlook for Food Consumption by Japanese Household: Considering the Impacts of Low

Birthrate, Aging and Altering Generation,

Tetsuro YAKUSHIJI (PRIMAFF)

✔ Agricultural Products Supply and Demand Simulation Model and Its Assumption,

Ninghui LI (IAED/CAAS)

✔ An Analysis on the Effect of the Distribution Control System for Tangerine,

Seong Bo KO(professer, Jeju National University)

11:10-11:20 Break Time

11:20-12:30 Agenda 1: Discussion

Moderator: Dong Yang WANG, Vice DG from China

Panelists: Suk Ho HAN(KREI)

Tomoo HIGUCHI(PRIMAFF)

Xiande LI(IAED/CAAS)

Q&A

12:30-14:30 Lunch

14:30-16:10 Agenda 2 Presentation: Green Growth in Agriculture and Rural

Communities

Moderator: Kiyoshi CHO, DG from Japan

✔ Implementation of Green Growth Strategy in Agriculture

Chang Gil KIM(KREI)

✔ Impacts of Climate Change on China's Agriculture,

Wei XIONG (Institute of Environment and Sustainable Development in Agriculture, CAAS)

✔ Building a Supply-Demand System for Eco-Friendly Crops in Jeju for Green Growth,

Seung Jin KANG(Jeju Development Institute)

✔ Current Situation on the Emissions Trading Scheme and Agricultural Sector in Japan,

Daisuke SAWAUCHI(PRIMAFF)

16:10-16:30 Break Time

16:30-17:50 Agenda 2: Discussion

Moderator: Kiyoshi CHO DG from Japan

Panelists: Shin Chan LEE(Jeju Agricultural Research & Extension Service)

Takeo TOMONO(PRIMAFF)

Suoping LI(IAED/CAAS)

Q&A

17:50-18:00 Closing

English Version

1. Outlook for Food Consumption by Japanese Household: Considering the Impacts of

Low Birthrate, Aging and Altering Generation / Tetsuro YAKUSHIJI ···························· 5

2. Agricultural Products Supply and Demand Simulation Model and Its Assumption

/ Li Ninghui ·························································································································· 19

3. An Analysis on the Effect of the Distribution Control System for Tangerine / Seong Bo KO ···················································································································· 37

4. Implementation of Green Growth Strategy in Agriculture / KIM Chang Gil ··············· 57

5. Impacts of Climate Change on China’s Agriculture / XIONG Wei ······························· 73

6. Building a Supply-Demand System for Eco-Friendly Crops in Jeju for Green Growth

/KANG Seung-Jin ··············································································································· 95

7. Current Situation on the Emissions Trading Scheme and Agricultural Sector in

Japan/Daisuke SAWAUCHI ······························································································ 113

Outlook for Food Consumption by Japanese Household: Considering the Impacts of Low Birthrate, Aging and Altering Generation

Tetsuro YAKUSHIJI

Policy Research Institute Ministry of Agriculture, Forestry and Fisheries

1. Introduction The falling birthrate and aging of the population are expected to progress.1 The demographic structure of Japan is such that persons aged 65 and over will comprise 30.5% of the population by 2025 (20.2% as of 2005) while the percentage of persons under age 14 will fall to 10.0% (13.8% as of 2005). Likewise, household composition is predicted to change with one-person households increasing 24.0% to comprise 36.0% of the total (29.5% as of 2005).2 In this way, the continued falling birthrate and aging coupled with the change in household composition is expected to have a considerable impact on food consumption in Japan. In this paper, we provide, under certain assumptions, an outlook for food consumption under the conditions of the falling birthrate and aging based on an analysis of changes in per-household food expenditures to date. 2. Basic idea behind the outlook In developing the outlook in this paper, we posit that food consumption for certain items at certain years in certain age groups is determined by the cohort effect, which differs according to birth year, the age effect, which is concurrent with aging, the period effect, which stems from the change in eras, consumption expenditure, and price. By analyzing these factors we examined what kind of impact these effects have had on consumption to date and what kind of impact they will have in the future. Conventionally, one ascertains these effects by looking at individuals3, but in this paper we use per-household data by age group of household head. Most food purchases in the household are not made individually by each household member but by a housewife or other household member who makes purchases for the entire household. With this in mind, we believe the use of per-household data is permissible. Therefore, the cohort effect and age effect referred to hereinafter are not effects for individual household members, but the cohort effect that stems from difference in the birth year of the household head for household to which other household members are party and the age effect concurrent with the aging of the household head. As a result, the age effect includes both changes in preference concurrent with ageing and changes in per-capita consumption in line with changes in family composition, such as the birth of children, their growth and their leaving the household as well as changes in lifestyle. In this sense, the age effect we refer to possess such qualities that it could also be called the life-stage effect4.

There are numerous analyses of food consumption that focus on the cohort effect. One of these is the compilation edited by Mori [3]which covers a wide array of issues—such as estimating individual consumption for several items per age group and problems with estimating the cohort effect—that concern this kind of analysis. More recent works include the analysis by Stewart et al. [4] to link the results of past cohort analyses to future outlooks. They examined the impact of the cohort effect on a vegetable consumption outlook in the United States and found that expenditures on household consumption for younger generations are declining and that expenditures on fresh vegetables will decline when these younger generations replace the older generations. While some works like this one conduct a thorough analysis of one specific item, it seems there are no works to date that have analyzed food consumption across the board. In this paper, we analyze all the items that comprise food expenses and offer an outlook for food consumption. The following analysis owes a great deal to Stewart et al. [4]. 3. Data and model (1) Data used In this paper we use data from the Family Income and Expenditure Survey, the National Survey of Family Income and Expenditure, and the Annual Report on Consumer Price Index, which are all issued by the Ministry of Internal Affairs and Communications, as well the Households Projections for Japan and the Population Projections for Japan, which are both issued by the National Institute of Population and Social Security Research. For two-or-more-person households we used the 21 years of data from 1987 to 2007 for expenditures by age group of household head (which did not include farmer, forester and fisher households through 2006) drawn from the Family Income and Expenditure Survey, and we used the 2005 data which includes farmer, forester and fisher households as the default5. For one-person households we used the five years of data from 1984 to 2004 for expenditures by sex and age group drawn from the National Survey of Family Income and Expenditure, a survey that is conducted every five years. We used the household number forecast issued in March 2008 and the population forecast issued in December 2006. In general, the older a household becomes the more it spends on items at high prices; therefore, we eliminated to the extent possible the expenditure differences caused by price differences among household head age groups by using average prices. For two-or-more-person households we calculate per-capita real expenditures by dividing figures for households by household member numbers for each household head age group, after having calculated real expenditures at 2005 prices using the Consumer Price Index. (2) Outlook model Since per-age group data are the average values for each group, we calculated our estimate

using the method of weighted least squares (WLS). We develop the following model to estimate coefficients:

Where, E(it) : Real expenditures per household member (For age group i and year t) (Same throughout)) D(it)2c : Cohort dummy D(it)3a : Age group dummy D(it)4p : Period dummy D(it)5h : Consumption expenditure coefficient dummy Y(it) : Per capita consumption expenditure

(per age group, per year) P(t) : Price (per year) e(it) : Error and βik is the coefficient to be estimated. In this equation, the second item on the right-hand side is the cohort effect, the third item is the age effect and the fourth item is the period effect. When one attempts to divide consumption by the cohort effect, age and the period effect, there is a relationship among birth year, age and period (i.e., year) of birth year + age = year; therefore, the problem of variables not being linearly independent has been pointed out6. To counter this problem, the forecited Stewart et al. [4] paper posits that persons born in or around the same year possess similar experiences (and exhibit similar behaviors). Based on this, they compiled cohort variables in five-year class interval, age variables in three-year class interval and period variables in two-year class interval in order to avoid this problem (The data used was raw data from Family Income and Expenditure Surveys conducted every three years between 1982 and 2003). We have avoided this problem by similar method although we do not have much freedom to set categories since the data used in this paper is aggregate data. Unlike data for two-or-more person households, data for one-person households is severely lacking, so we used the results from the National Survey of Family Income and Expenditure which can be used for age groups in 10-year class interval. Since this survey is only conducted every five years, the data we used was the data collected every five years between 1984 and 2004; however, per-gender data for one-person households could be used. For this reason, we developed an outlook with 60 samples over both genders x 6 age groups x 5 years. We felt the sample size was too small for one-person households to seek the coefficients for consumption expenditures and price simultaneously with other coefficients as we did with two-or-more-person households. As such, we used the coefficients for two-or-more-person households for these. More specifically, we calculated the consumption expenditure elasticity

)()()()(

)()()()(

)(

)log(

itn

2h

t6

itith5h551

n

2p

itp4p4

n

2a

ita3a3

n

2c

itc2c21

it

ePYD

DDDE

h

pac

+⋅+⋅⋅+

+⋅+⋅+⋅+=

∑

∑∑∑

=

===

βββ

ββββ

　　　　　

　

(per age group) and price elasticity from the results of estimate for two-or-more-person households, and using these, we estimated the other coefficients as independent variables by removing the elements pertaining to consumption expenditure and price from log (E(it)). Here, D(it)7g is the female dummy and is defined for each age group g (three groupings).ηYh is the consumption expenditure elasticity for age group h, ηp is the price elasticity and D(it)5h is the age group dummy. The number of independent variables for two-or-more-person households totaled 34: 13 for cohorts, nine for age, six for period, one for consumption expenditure, four for consumption coefficients and one for price. The number of samples for this was 10 age groups x 21 years = 210. For one-person households there were a total of 18 independent variables for a sample size of 60: three for female, six for cohorts, five for age and four for period (See Table 1 for details on variables used). The sample size for two-or-more-person households under the age of 24 was small, so there are large fluctuations according to year. Given this, we defined an outlier dummy for the data which were thought to clearly represent outliers in order to prevent an adverse impact on the coefficient estimation. Likewise for one-person households, we incorporated an outlier dummy for data thought to be outliers as determined from the preceding or following year.

)()()()(

)()()()()( )log()log()log(

itn

2p

itp4p4

n

2a

ita3a3

n

2c

itc2c2

n

1g

itg7g1

n

1h

tP

itith5Yh

it

eDDD

DPYDE

pac

gh

+⋅+⋅+⋅

+⋅+=⎟⎟⎠

⎞⎜⎜⎝

⎛⋅+⋅⋅−

∑∑∑

∑∑

===

==

　　　　　　　　　　　　　　　

７

βββ

ββηη

Model for two-or-more-personhouseholds

Model for one-person households

Female dummy None ～39 years old40～5960～

(Number of variables) 3Cohort dummy ～1922 (*) ～1927 (*)(Birth year dummy) 1923～27 1928～37

1928～32 1938～471933～37 1948～571938～42 1958～671943～47 1968～771948～52 1978～871953～571958～621963～671968～721973～771978～821983～87

(Number of variables) 13 6Age dummy ～24 years old (*) ～29 years old (*)

25～29 30～3930～34 40～4935～39 50～5940～44 60～6945～49 70～50～5455～5960～6465～

(Number of variables) 9 5Period dummy 1987～89 (*) 1984 (*)

1990～92 19891993～95 19941996～98 19991999～01 20042002～042005～07

(Number of variables) 6 4Consumption expenditure (～29 years old) (**)

(Number of variables) 1～29 years old (*)30～3940～4950～5960～

(Number of variables) 4 0

PriceCommon coefficient for all the agegroups

Unnecessary because elasticitiesestimated by the model for two-or-more-person households are used.

(Number of variables) 1 0

Number of independentvariables

34 18

Table 1. List of variables

* Excluded from variables** The coefficient of this variable represents the group that is less than 29 years old because this age group is excluded from the slope dummy variables.

Unnecessary because elasticitiesestimated by the model for two-or-more-person households are used.　　～29 years old　　30～39　　40～49　　50～59　　60～69 （same to 60～）　　70～ （same to 60～)

Slope dummy ofconsumption expenditure

(3) Estimating per-capita future real consumption Regarding the future outlook, we sought E , the per-household member per-age group expenditures for the estimate year, from the exogenously given D(it)2c，D(it)3a，D(it)4p，D(it)5h，Y(it) and P(t). Then, we multiplied this by the future average number of household members per household and number of households7 for each household head age group to calculate overall expenditures. When we exogenously determine D(it)2c, we equated the cohort effect for the cohort which will newly enter the lowest age group going forward with the current lowest age group.

Regarding the coefficient for period effect D(it)4p, when the factor clearly exhibited an increasing or decreasing trend, we altered the future coefficient accordingly. More specifically, we examined the trend in the coefficient for period effect, and when no specific trends could be found, we set the coefficient for the most recent year. When a specific trend could be found, as a rule we adjusted the coefficient by the average annual increase or decrease for the coefficients for the past two periods (six years for two-or-more-person households; 10 years for one-person households) over the 10 years between 2005 and 2015, and we adjusted for the subsequent 10 years by halving that quantity.8

For consumption expenditures Y, we calculated changes in per-capita consumption expenditures for future household surveys from the GDP growth rate and population growth rate for Japan used by the OECD [5]9. The average per-capita real GDP growth rate that we used was then 1.6% for 2005 ~ 2015 and 1.5% for 2015 ~ 2025. Price P was set at the 2005 level. Since we do not think that consumption is dependent upon the reasons listed above, we set values for school lunch (only for two-or-more-person households) proportionate to changes in the number of pupils. 4. The cohort effect, age effect and period effect in several items

We looked at uncooked rice and cooked food with rice, bread or noodles for two-or-more-person households and observed the coefficients for the estimated cohort dummy, age dummy and period dummy to examine the impact of cohort effect, age effect and period effect on past expenditures. Our analysis is as follows. Uncooked rice is a household cooking ingredient and an example of an item for which demand has fallen. Cooked food with rice, bread or noodles is a home meal replacement and an example of an item for which demand has risen. Here we set the value of the variables removed from the dummy variables at zero and have included them in the graph. Some of the estimated coefficients have low t-statistics. In the graph, coefficients with a t-statistic of two or higher are highlighted. In addition to these coefficient estimates generated by all variables, we used the backward elimination method to reduce the number of variables and estimate significant coefficients. We did not use them in the projection, but marked coefficients with a t-statistic of two or higher with an X.

(1) Rice Since the generation born in the 1940s, the cohort effect has decreased continuously and significantly. The age effect from the 30s through the 50s is high, but falls for the 60s and higher. We believe this reflects changes in per-household member consumption concurrent with changes in the age composition of household numbers due to the birth of children, their growth and their leaving the household. The period effect has decreased continuously and significantly (See Figures 1－1－1 ~ 1－1－3).

Given these results, both the cohort effect and the period effect (if this trend continues) are expected to move to reduce future per-capita consumption.

(2) Cooked food with rice, bread or noodles The cohort effect for cohorts born after around 1935 increases as birth year’s decrease,

reaching its peak around 1955. After that it stabilizes (and coefficients to this point are significant) and makes another upturn around birth year 1980. The age effect is highest for 50 year olds. The period effect is on an upward trend and is significant (See Figures 1－2－1 ~1－2－3).

In the future, it is highly likely that the cohort effect will contribute to an increase in consumption, and the same applies to the period effect if this trend continues.

Figure 1-1-1. Cohort effect - 001 Rice

-1.0-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.10.00.1

～1922

1923～

27

1928～

32

1933～

37

1938～

42

1943～

47

1948～

52

1953～

57

1958～

62

1963～

67

1968～

72

1973～

77

1978～

82

1983～

87

Birth year of household head

Figure 1-1-2. Age effect - 001 Rice

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

～24

25～

29

30～

34

35～

39

40～

44

45～

49

50～

54

55～

59

60～

64

65～

Age of household head

Figure 1-1-3. Period effect - 001 Rice

-0.4

-0.3

-0.2

-0.1

0.0

1987～

89

1990～

92

1993～

95

1996～

98

1999～

01

2002～

04

2005～

07

Year

5. Results of the future outlook (1) Outlook for 30 food items

Figures 2－1 ~ 2－3 show the past change rate (1990 – 2005 (converted to a 20-year change rate)) and the future change rate (2005 – 2025) for overall real consumption of all of the 30 food items. Data is shown for two-or-more-person households, one-person households, and all households (two-or-more-person households + one-person households). 1) Two-or-more-person households Due to the decrease in households and household members per household, total food

expenditures will fall 11.8%. Looking at each item, expenditures for almost all items will fall with expenditures for six on this rise and expenditures for 24 decreasing. Amidst this trend, expenditures for cooked food with rice, bread or noodles will post a major increase. Expenditures for beverages will also increase. Among items on which expenditures will fall, the decline for rice, fish and shellfish, raw meat, fresh vegetables, fresh fruits and other fresh foods will be steep, while the rate of decline for bread, seasonings, oils and fats, processed meat, soybean products, processed fruits and other processed items and for meals outside the home will be small. While it is not shown here, when one examines per-capita expenditures,

Figure 1-2-1. Cohort effect - 023 Cookedfood with rice, bread or noodles

-0.1

0.0

0.1

0.2

0.3

0.4

～1922

1923～

27

1928～

32

1933～

37

1938～

42

1943～

47

1948～

52

1953～

57

1958～

62

1963～

67

1968～

72

1973～

77

1978～

82

1983～

87

Birth year of household head

Figure 1-2-2. Age effect - 023 Cooked foodwith rice, bread or noodles

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

～24

25～

29

30～

34

35～

39

40～

44

45～

49

50～

54

55～

59

60～

64

65～

Age of household head

Figure 1-2-3. Period effect - 023 Cookedfood with rice, bread or noodles

0.00.10.20.30.40.50.60.7

1987～

89

1990～

92

1993～

95

1996～

98

1999～

01

2002～

04

2005～

07

Year

expenditures for eating out, other processed foods, bread, seasoning and oils and fats are expected to increase. A shift from fresh items to processed items and from meals inside the home to eating out and home meal-replacement is forecast (Figure 2－1).

2) One-person households Due to the increase in households, total food expenditures will grow 33.7%. Looking at each item, expenditures for almost all items will increase with expenditures for 24 on this rise and expenditures for five falling. Among items on which expenditures will rise, the rate of increase for processed fruits, other cooked foods, oils and fats, tea, soybean products, seasonings, cooked food with rice, bread or noodles, dairy products and other processed items will be high, the rate of increase for rice, fresh vegetables, fish and shellfish and other fresh items will be low. Despite this situation, expenditures for eating out will diminish. A shift from fresh items to processed items is predicted for one-person households as well, but unlike two-or-more-person households, a shift from eating out to home meal replacement is also anticipated (Figure2－2). 3) All households (Two-or-more-person households + one-person households)

Total food expenditures will remain almost the same, falling 1.9%. Looking at each item, expenditures 14 items will increase while expenditures for 16 items will fall—an almost even split. Major increases are forecast for cooked food with rice, bread or noodles, other cooked foods and beverages. The past and future directions of change are the same for almost every item except eating out which will shift from an increase to a decline. This is because they are predicted to fall in one-person household despite the rise in its number. Due to falling household numbers and number of members per household, expenditures for many fresh food items will fall. Expenditures for many processed items are forecast to rise because of the increase in

Figure 2-1. Expenditure change (%)(Two-or-more-person households)

School lunch

Eating out

Alcoholic beverages

Other beveragesCoffee & cocoa

Tea

Other cooked food

Cooked food with rice,bread or noodles

Cakes & candies

SeasoningsOils & fats

Processed fruits

Fresh fruits

Other processedvegetables & seaweeds

Soybean products

Dried vegetables &seaweeds

Fresh vegetables

Eggs

Dairy products

Fresh milk

Processed meat

Raw meat

Other processed fish

Fish-paste products

Salted & dried fish

Raw fish & shellfish

Other cereals

Noodles

Bread

Rice

Food

-60

-50

-40

-30

-20

-10

0

10

20

30

40

-60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100

1990→2005 Change rate

2005→

2025 C

han

ge r

ate

Note: 1990→2005 Change rates have been converted into 20-year change rates.

one-person households. A shift from fresh items to processed items and from at home cooking to purchasing cooked food or home meal replacement is expected to progress. Our diet will be much more dependent on food industry (Figure 2－3).

Figure 3 shows the ratio of each food item to total food expenditures. To avoid complication, the 30 food types have been re-grouped into 12 categories. The six items on the bottom of the legend (from Cereals up to Fruits) are those for which ratios will continue to contract, while the next four items (from Oils, fats and seasonings up to Beverages) are those for which ratios will increase. The ratio for alcoholic beverages hardly changes, while that for meals outside the home falls. Cooked food and meals outside the home comprised 34.1% of food and drink expenses in 2005, and this will rise to 37.4% in 2025. Meanwhile, the ratios for fresh items such as rice, fish and shellfish, fresh meat, eggs, fresh vegetables and fresh fruits continue to fall drastically over the years: 32.1% in 1990, 31.0% in 1995, 29.1% in 2000, 26.8% in 2005, 23.5% in 2015 and 21.3% in 2025.

Figure 2-3. Expenditure change (%)(All households)

School lunch

Eating out

Alcoholic beverages

Other beverages

Coffee & cocoa

Tea

Other cooked food

Cooked food with rice,bread or noodles

Cakes & candies

Seasonings

Oils & fats

Processed fruits

Fresh fruits

Other processedvegetables & seaweeds

Soybean products

Dried vegetables &seaweeds

Fresh vegetablesEggs

Dairy products

Fresh milk

Processed meat

Raw meat

Other processed fish

Fish-paste products

Salted & dried fish

Raw fish & shellfish

Other cereals

Noodles

Bread

Rice

Food

-50

-30

-10

10

30

50

-50 -30 -10 10 30 50 70 90 110 130

1990→2005 Change rate

2005→

2025 C

han

ge r

ate

Note: 1990→2005 Change rates have been converted into 20-year change rates.

(2) Outlook by household head age group and household type Looking at changes in expenditure ratios as a part of overall food expenditures by household head age group (Figure 4), the expenditure ratio for households with household heads aged 60 or older will rise from 37.0% in 2005 to 47.5% in 2025, comprising nearly half of all expenditures.

Figure 3. Composition ratio of food expenditureby commodity group(All households) (%)

7.98.28.68.99.39.77.2

8.09.410.211.511.7 6.2

6.77.58.39.1

9.53.9

4.14.54.7

4.54.210.1

10.511.111.5

11.411.73.03.4

3.84.04.04.5

4.54.34.03.7

3.33.17.47.3

6.76.36.56.7 16.614.7

12.010.99.78.3

7.76.95.64.64.13.8

4.84.74.84.84.84.8

20.821.222.122.121.721.9

0

10

20

30

40

50

60

70

80

90

100

1990 1995 2000 2005 2015 2025

Meals outsidethe home

Alcoholic beverages

Beverages

Cooked food

Cakes & candies

Oils, fats & seasonings

Fruits

Vegetables &seaweeds

Dairy products& eggs

Meat

Fish & shellfish

Cereals

Note: 2005 price. Totaled in 12 commodity groups.

Figure 4. Composition ratio of food expenditureby age group of household heads

(All households) (%)

9.7 9.6 9.3 8.1 6.6 6.5

17.0 14.6 14.5 14.3 12.0 10.2

29.6 27.1

21.1 17.7

18.5 16.2

22.6 23.1

25.0 23.0

17.9 19.7

21.1 25.6 30.0 37.0

45.0 47.5

0

10

20

30

40

50

60

70

80

90

100

1990 1995 2000 2005 2015 2025

60～

50～59

40～49

30～39

～29

Note: 2005 price.

Looking at expenditure ratios as a part of overall food expenditures for one-person households versus two-or-more-person households (Figure 5), ratios for one-person households will rise from 21.7% in 205 to 29.6% in 2025, comprising nearly 30% of all expenditures.

Figure 6 shows the per-item consumption ratios for households with household heads aged 60 or older. Consumption ratios for almost all items rise in households with household heads aged 60 or older. One can see higher ratios for fruits (fresh fruits and processed fruits); seafood (raw fish and shellfish, salted and dried fish, fish-paste products and other processed fish); and dried vegetables and seaweeds and lower ratios for eating out; beverages (tea, coffee and cocoa and other beverages); cooked food with rice, bread or noodles; bread; noodles and meat (raw and processed).

Figure 6. Proportion of expenditure by households with heads aged 60 or older (All household) (%)

Eating out

Alcoholic beverages

Other beverages

Coffee & cocoa

Tea

Other cooked foodCooked food with rice,bread or noodles

Cakes & candies

Seasonings

Oils & fats

Processed fruits

Fresh fruitsOther processed

vegetables & seaweeds

Soybean products

Dried vegetables &seaweeds

Fresh vegetables

Eggs

Dairy products Fresh milk

Processed meat

Raw meat

Other processed fish

Fish-paste products

Salted & dried fish

Raw fish & shellfish

Other cereals

NoodlesBread

Rice

Food

20

30

40

50

60

70

80

20 25 30 35 40 45 50 55 60

2005

2025

Figure 5. Composition ratio of food expenditureby age group of household head,by family type of household (%)

1.81.82.33.23.33.17.08.110.3

11.211.814.5

11.914.314.8

18.524.5

27.614.7

14.719.9

22.2

21.120.9 35.0

35.5

30.925.2

21.818.3

4.7

4.85.86.26.3

6.6

3.23.9

4.03.32.72.4

4.34.2

2.92.62.61.9

4.93.2

3.02.82.01.7 12.59.4

6.14.83.82.8

0

10

20

30

40

50

60

70

80

90

100

1990 1995 2000 2005 2015 2025

60～

50～59

40～49

30～39

～29

60～

50～59

40～49

30～39

～29

Note: 2005 price.

One-person households

Two-or-more-personhouseholds

21.7 25.5 29.6

Figure 7 shows the per-item consumption ratios for one-person households. Excluding eating out, ratios in 2025 will increase over 2005 levels. Items with high ratios in 2005 that will rise higher include tea, coffee and cocoa and cooked food with rice, bread or noodles. Items with low ratios in 2005 that will not rise significantly include meat (raw and processed), salted and dried fish and eggs.

6. Conclusion In this paper, we examined the cohort effect, age effect and period effect as factors that have an impact on per-capita consumption in household for the expenditures on 30 food items in the Family Income and Expenditure Survey. In addition to price and consumption expenditures, we analyzed these effects on past data, and based on this, we provided an outlook for future consumption given certain assumptions.

By offering a future outlook that assumes a measure of growth in consumption expenditures, our findings have shown a shift from fresh items to processed items, from at-home cooking to purchasing cooked food or home meal replacement, so that the externalization of food, more specifically the dependency of our diet on food industries, will progress. Let us briefly touch on the relationship between these findings and consumers’ recent return to cooking at home. This point was mentioned in the Annual Report on Food, Agriculture and Rural Areas in Japan FY2008, but we did not examine it in this paper. If this is a reflection of the recent economic situation, then trends would move toward another dependency on food industries concurrent with a long-term economic recovery. As mentioned earlier, the outlook in this paper was developed based on a certain measure of GDP growth. We should state that this is one reason why, in our findings, expenditures on processed items and cooked food—both which have high consumption expenditure elasticities—will rise sharply, contributing to continuing

Figure 7. Proportion of expenditure by one-person households (%)

Eating out

Alcoholic beverages

Other beverages

Coffee & cocoaTea

Other cooked food Cooked food with rice,bread or noodles

Cakes & candiesSeasonings

Oils & fats

Processed fruits

Fresh fruits

Other processedvegetables & seaweeds

Soybean products

Dried vegetables &seaweeds

Fresh vegetables

Eggs

Dairy products

Fresh milk

Processed meat

Raw meat

Other processed fish

Fish-paste products

Salted & dried fish

Raw fish & shellfish

Other cereals

Noodles

Bread

Rice

Food

5

10

15

20

25

30

35

40

45

5 10 15 20 25 30 35 40 45

2005

2025

externalization of food. In closing, the estimates in this paper are for items purchased as part of household budgets, so we did not give any consideration to demand for agricultural raw materials. For example, we predict a steep decline in the consumption of raw items purchased as part of household budgets, but the consumption of agricultural raw materials will increase indirectly by way of processed items purchased and the shift to home meal replacement. When viewed through the lens of agricultural raw materials, direct consumption in household will fall, but the potential exists for a large increase in demand for raw materials for processed goods. While the extent of this change is unclear, we must point out that, in this paper, as consumption of fresh items by household declines, overall demand for agricultural raw materials will not fall.

Literature Cited

[1] National Institute of Population and Social Security Research. “Population Projections for Japan”

(December 2006 estimates). Available at http://www.ipss.go.jp/syoushika/tohkei/suikei07/index.asp (Downloaded November, 21, 2008)

[2] National Institute of Population and Social Security Research. “Households Projections for Japan” (March 2008 estimates). Available at http://www.ipss.go.jp/pp-ajsetai/j/HPRJ2008/t-page.asp (Downloaded November, 21, 2008)

[3] Hiroshi Mori, ed., (2001). Cohort Analysis of Japanese Food Consumption—New and Old Generations, August 2001, Senshu University Press.

[4] H. Stewart and N. Blisard (2008), “Are Younger Cohorts Demanding Less Fresh Vegetables?”, Review of Agricultural Economics, Vol. 30, No. 1, Spring 2008.

[5] OECD (2008), OECD-FAO Agricultural Outlook 2008-2017. 1 Based on National Institute of Population and Social Security Research estimates for medium variant for fertility

(and medium variant for mortality) in [1]. 2 Based on [2].

3 Mori, ed. [3] contains individual consumption estimates by age for fruits and several other items.

4 Based on a suggestion from Professor Yukihiko Uehara, Meiji University Global Business Graduate School. 5 Individual data, not aggregate data, is used in Stewart et al. [4]. 6 Identification issues such as these are discussed in detail in Hirohiko Asano’s “Comparative Methodological

Study of Cohort Analysis” (in Mori, ed. [3]). 7 In the National Institute of Population and Social Security Research’s household number estimates, estimates are

given for household numbers per age group of household head but not for household member numbers. For this reason, we separately estimated household member numbers by using population estimates for two and three-generation households per family types in the 2005 Population Census.

8 We halved the change rate since it is difficult to assume that change over the past six to 10 years will continue at the same rate for 20 years.

9 There is a major discrepancy between the change rate for household consumption expenditures derived from National Accounts and the change rate for consumption expenditures (for two-or-more-person households) derived from the Family Income and Expenditure Survey. For this reason, based on the relationship between past per-capita real GDP growth and the Family Income and Expenditure Survey-based per-capita real consumption expenditure change rate, we converted the real GDP growth rate into the Family Income and Expenditure Survey-based real consumption expenditure change rate, thereby making it exogenous.

Agricultural Products Supply and Demand Simulation Model and Its Assumption

LI Ninghui

Institute of Agricultural Economics Chinese Academy of Agricultural Sciences

Chinese food security is important not only for China’s economy development and Chinese

livelihood but also for the whole world’s food security, which makes it necessary to simulate and project the variation of production, consumption, and trade of agricultural products, project In order to analyze various agricultural economic behaviors, relations, phenomena, effects on whole national economy, and future development trends, it is necessary to find factors affecting economy simultaneously and to build some dynamic economic model system for which some new theories and methods are appealed. It is this consideration that elicit CAPSiM (CCAP’s Agricultural Policy Simulation Model)..

Theoretical structure of CAPSiM

CAPSiM is driven either by endogenous or by exogenous determinants of supply and demand. Supply equations, which are decomposed by area and yield for crops and output for meat and other products, allow producer own price and cross market responses, as well as the effects of shifts in technology stock on agriculture, irrigation stock, ratio of erosion area to total land area, ratio of salinity area to cultivate area, yield change due to exogenous shock of climate, and yield change due to other exogenous shock.

Demand equations, which are decomposed by urban and rural, allow consumer own-price and cross market responses, as well as the effects of shifts in income, population level, market development and other shocks.

The general framework is presented in the following figure.

O u tp ut p ri c e sI n p u t p ri ce sI n v es tm e n t：－ R& D－ I rri g a ti o n

W e at he rO th ers

Are a

Yi el dP ro duc tion

T o ta lSu p p ly

St o ck

Im por t

E x po r t

Pri ce

T ot a lD ema n d

F o o d

F eed

S e ed / O th ers

F ig u re III. 1 C APS iM F ram ew o rk

L i v e s to c k

Mac ro P o l ic y

I n c o m eP o p u l a ti o nHa b i t ch a n g eP r ic e sO th ers

The theoretical models of CAPSiM are given below.

Let: X̂ = dX/X (percentage change in X) in the following equations.

1. Domestic Production 1.1. Crop Production

Crop production is the product of harvested area and yield. Crop harvested area is a Cobb-Douglas function of the crop’s own price and other crops’ prices, and responds to the exogenous shocks from climate, policy and the other sources. Yield is a Cobb-Douglas function of the crop’s own price, agriculture technology stock and irrigation stock, and responds to the exogenous shocks from climate and other sources. All parameters used in these models are econometrically estimated.1

Area: log Ait = aAi0 +ΣjbAij(log pSjt) Yield: log Yit = aYi0 +ΣjbYij(log pSjt) + cilog Rt + kilog It + gilog (Erosiont) + hilog (Salinityt) Production: QSit = Ait * Yit Variation Relationship:

SitQ̂ = it + itŶ + (Z

A1i(t-1) + ZA2it + ZA3it + ZY1it + ZY2it ).

Where: A: Crop harvested area. ZA1: percentage change in area due to exogenous shock of climate. ZA2: percentage change in area due to exogenous shock of policy. ZA3: percentage change in area due to other exogenous shock. pS: prices of output and input for producer. Y: crop yield per hectare. R: agriculture technology stock. I: irrigation stock. Erosion: ratio of erosion area to total land area. Salinity: ratio of salinity area to cultivate area. ZY1: percentage change in yield due to exogenous shock of climate. ZY2: percentage change in yield due to other exogenous shock. i: index crop, including: rice, wheat, maize, sweet potato, potato, other coarse grains,

soybean, cotton, oil crop, sugar crop, vegetable, and other crops. j: index crop output and input, including: rice, wheat, maize, sweet-potato, potato, other

1 The parameters of these two models (1998) are given in Appendix 2 Table 1 and Table 2.

coarse grains, soybean, cotton, oil crop, sugar crop, vegetable, and other crops, fertilizer, labor, and land.

Constraints: When i>j, bAij = bAji*Aj/Ai because (dAi/dpjS)/(dAj/dpiS)=piS/pjS. It means that the area's

marginal rate of substitution between any two crops must be equal to their price ratio.

∑j

Aijb = 0 because of homogenous of degree zero. It means that an equal percentage

change in all prices leads to no change in the crop area response. When i ≠ j, bYij = 0, it means that crop yield responds to the change in its own price

only. Note: • In this crop production model, if CAPSiM runs in endogenous mode and if Ait is fixed

exogenously, at least one of the three shocks, ZA1i(t-1), ZA2it and ZA3it, must be endogenized. Similarly, if Yit is fixed exogenously, at least one of the two shocks, ZY1it and ZY2it, must be endogenized.

• When i>j, bAij varies yearly due to the first constraint. 1.2. Livestock Production

Livestock production is a Cobb-Douglas function of prices of output and input for producer and responds to the exogenous shocks from disease and the other sources.2

Production: log qit = aqi0 +Σjbqij(log pSjt) Variation Relationship:

itq̂ = shockswithoutitq _|ˆ + Zq1

it + Zq2it

Where: q: total livestock production. Zq1: livestock production change due to exogenous shock of disease. ZY2: livestock production change due to other exogenous shocks.

pS: prices of output and input for producer. i: index meat production, including: pork, beef, mutton, poultry, egg, milk, fish.

j: index meat and input, including: pork, beef, mutton, poultry, egg, milk, fish, maize and labor.

Constraints: When i ≠ j, bqij = 0, it means that livestock production responds to the change in its own

price only Production by Mode: In China, livestock is produced in three production modes: backyard, specialized household,

and commercial intensive production. In order to analyze the technological coefficients of input

2 Parameters of this model (1998) are given in Appendix 2 Table 3.

and output in livestock production, the total livestock production is decomposed into three parts according to these three modes. The share of output in each production mode is subject to be changed from year to year.3

qikt = θikt * qit θikt = θik(t-1) + γik Σkθikt = 1 Where: θ: share of each production mode. γik: annual change in θ. k: index production mode, including: backyard, specialized household, and commercial

intensive production. 2. Domestic Demand

On the demand side, the changes in urban economy have made urban consumers almost

entirely depend on markets for their consumption need. In this sector, prices and income changes will likely be the fundamental forces driving change in consumption patterns.

Rural residents live in a very different environment from their urban counterparts and exhibit a different kind of demand behavior. Taking this fact into account, the demand models for urban and rural areas are given separately, which differ in the inclusion of the food market development index.

Among the domestic demand models, grain consumption is divided into two parts: grain that is directly consumed for food and grain that is fed to animals and consumed indirectly which is inputted from underlying demand equations for pork, beef, mutton, poultry, fish, eggs and milk.

Similar to supply side, this part of the analysis uses econometrically estimated parameters. Demand parameters are estimated using an Almost Ideal Demand System framework and based on household survey data. Elasticities of expenditure are estimated to vary according to the level of income. 2.1. Food Demand

Food demand is a Cobb-Douglas function of consumer prices, per capita income and food

market development index in rural area. At first, per capita demands in rural and urban areas are calculated separately, then, national per capita demand is calculated by using rural and urban population as weights.4

log dRit = aRDi0 +ΣjbRij(log pDjt) + eRilog YRt + milog (MKTt) log dUit = aUDi0 +ΣjbUij(log pDjt) + eUilog YUt dit = θRtdRit + θUtdUit Dit = dit * Popt Where:

3 Coefficients are given in Appendix 2 Table 4. 4 Parameters of this model are given in Appendix 2 Table 5 (Urban) and Table 6 (Rural).

dR, dU: per capita demand in rural, urban. d: national per capita demand. D: national total demand. pD: consumer price. YR, YU,: per capita income in rural, urban. MKT: food market development index. bR, bU: price elasticity matrix of demand in rural, urban. eR, eU: income elasticity vector of demand in rural, urban. m: market development elasticity of demand in rural. θR, θU: rural share, urban share in the total population. Pop: total population i: index non-livestock food, including: rice, wheat, maize, sweet potato, potato, other coarse

grain, soybean, sugar, oil, fruit, vegetable, and other food, pork, beef, mutton, poultry, egg, milk, and fish.

Constraints: When i>j, bRij = Expenditure Sharej * (bRji / Expenditure Sharei + Real incomej - Market

developmenti).

Cournotj = ∑i

Rijb *Expenditure Sharei + Expenditure Sharej. Walras' law. (The

properties of Cornot and Engel aggregation). It means that total expenditure cannot change in response to a change in prices.

The constraints on bUij are the same as those on bRij except there is no Market Development in the first constraint.

Note: • When i>j, bRij and bUij vary yearly due to the first constraint.

2.2. Total Feed Demand

In this model, feed demand in different grain is obtained by tracing the quantity of grain

consumed by different livestock produced in different production modes. Once the demand for meat and other animal products is known, the implied feed demand is calculated by applying a set of feed conversion ratios. Then total feed demand is calculated by using the grain’s share as weight. The feed/meat ratio, efficient gain in feeding livestock and grain i’s share in total feed grain are estimated from other surveys out of CAPSiM. The feeding efficiency of hogs is expected to increase slightly over time. Meat products are assumed to be produced in China and to be sufficient to satisfy demand. The alternative assumption of net import of meat is also made to investigate the effects of meat imports on the demand for total grain.5

DFeedGjt = Σk(1 + δjkt)βjkθjktqjt θjkt = θjk(t-1) + γjk DFeedGt =ΣjDFeedGjt

5 θ, γ, β and δ are given in Appendix 2 Table 4, f and rf are given in Appendix 2 Table 7.

DFeedGit =ΣifitDFeedGt fit = (1 + rf)tfi(t-1) Where: DFeedG: Total Feed Demand. θ: share of each production mode.

γik: annual change in θ. β: Feed/meat ratio. δ: Efficient gain in feeding livestock. f: grain i’s share of total feed grain. rf: annual growth rate of f. i: index individual grain and other feed, including: rice, wheat, maize, sweet potato, other

coarse grain, and soybean. j: index meat product, including: pork, beef, mutton, poultry, egg, milk, and fish. k: index production mode, including: backyard, specialized household, and commercial

intensive (company) production. 2.3. Other Grain Demand

Other grain demand includes grain used as seed, grain used in industry and grain wasted due

to post-harvest loss. Quantity of grain used as seed depends on the crop-harvested area. Quantity of grain used in industry this year is calculated based on that used last year. Quantity of grain wasted is calculated as share of crop production.6

DSeedit = (1 + βSt)tdSeedi(t-1)Ait DIndit = (1 + βIi)tDIndi(t-1) DWasteit = (1 + βWi)tdWastei(t-1)QSit Where: A: Crop harvested area. DSeed: Seed grain use. DInd: Industrial grain use. DWaste: Grain post-harvest loss. dSeed: Seed grain use (kg) per hectare. dWaste: Waste (loss) as share of production. βS: annual growth of seed use per ha. βI: annual growth of industrial grain use. βW: annual growth of post-harvest loss. i: index individual grain, including: rice, wheat, maize, sweet potato, potato, other coarse

grain, soybean.

6 Coefficients of this model are given in Appendix 2 Table 8.

2.4. Total Grain Demand Total grain demand is the summation of total food demand, total feed demand, seed use,

industry use and total wasted which exhaust total grain consumption. DGit = DFoodGit + DFeedGit + DSeedit + DIndit + DWasteit Where: DG: Total Grain Demand. DFoodG: Total Food Grain Demand. i: index individual grain, including: rice, wheat, maize, sweet potato, potato, other coarse

grain, soybean. 3. Grain Stock

Grain stock this year is inferred by referencing the increase in gain demand and grain stock

last year as well as the effect of change in price on grain stock. BGit = Bstocki(t-1) (1 + ΨDGit / DGi(t-1)) –ΨBstocki(t-1) + LpDit Where: BG: Grain stock. pD: consumer price. L: marginal change in grain stock due to grain price change.7 Ψ = 0 if long-term grain stock is constant over time. That is, BGit - Bstocki(t-1) = LpDit, implying

the change in stock is a linear function of consumer price. It is this formula that is used in this study since we are doing long-term simulation and projection.

Ψ = 1 if the proportion of grain stock to consumption is constant over time. That is, BGit / DGit - Bstocki(t-1) / DGi(t-1) = LpDit, implying the change in the ratio of stock to consumption is a linear function of consumer price. This formula is used in short-term (usually less than 5 years) simulation and projection.

In terms of this formula, the stock change will be directly affected by domestic consumer price, and indirectly affected by world market price through trade because domestic consumer price will be affected by world market price via trade model. Because the value of L is negative, if the consumer price keeps decrease, the stock will keep increase (in the case of Ψ = 1, if the stock grows faster than consumption). In this case, according to the government stock policy, we give a maximum stock for each commodity (30% of each product production). If the stock grows so that the stock level is greater than the maximum, the export will be forced to increase exogenously while the stock level will be maintained unchanged. In the opposite case, the minimum stock level is set at 5% of production. In this case, if the stock decreases due to the increase in consumer price so that the stock level is less than the minimum, the import will be forced to increase while the stock level will be maintain unchanged.

7 L: rice –0.19, wheat –0.2, maize –0.3, sweet potato –0.3, potato –0.3, other grains –0.2, soybean –0.35.

Trade Behavior After estimating the changes in production of and demand for agriculture products by using the models given above, we can get the percentage changes in import and export by decomposition equation derived from constant elasticity of substitution (CES) mechanism, in response to the percentage changes in their prices domestically and abroad. Consider the two components case, where the elasticity of substitution is defined as the percentage changes in the ratio of the two cost-minimizing component demands, given a 1 percent change in the inverse of their price ration:

σ ≡ )./̂/()/̂( 1221 ppqq (V.1)

For larger values of σ, the rate of change in the quantity ratio exceeds the rate of change in the price ratio and the cost share of the component that becomes more expensive actually falls. Expressing equation (V.1) in percentage change form, we obtain:

).ˆˆ()ˆˆ( 1221 ppqq −=− σ (V.2)

CES functional form gives the following relationship between the changes in quantities demanded of components and the composite good:

,ˆ)1(ˆˆ 2111 qqq θθ −+= (V.3)

where θ1 is the cost share of component 1 and (1-θ1) is the cost share of component 2. Solving

for 2q̂ gives:

),1/()ˆˆ(ˆ 1112 θθ −−= qqq (V.4)

which may be substituted into (V.2) to yield:

.ˆ)ˆˆ()1(ˆ 1211 qppq +−−= σθ (V.5)

Note that this conditional demand equation is homogeneous of degree zero in price, and the compensated cross-price elasticity of demand is equal to (1-θ1)σ. CES functional form also gives the following relationship between the changes in prices of components and the composite good:

.ˆ)1(ˆˆ 2111 ppp θθ −+= (V.6)

First we solve for 2p̂ as a function of 1p̂ and p̂ , then substitute this to (V.5) to obtain:

.ˆ)ˆˆ(ˆ 11 qppq +−= σ (V.7)

Note that the form of equation (V.7) is unchanged when the number of components increases

beyond two. This equation decomposes the change in the derived demand, 1q̂ , into two parts.

The first is the substitution effect. It is the product of the constant elasticity of substitution and the percentage changes in the ration of the composite price to the price of component 1. The

second is the expansion effect. Owing to constant return to scale, this is simply an equi-proportionate relationship between composite and component. 4. Trade

In this trade model, FOB and CIF prices are first exchanged into domestic currency. After that,

they are transformed into domestic market prices at the national level by deducting producer subsidy expenditure. Then, the percentage changes in the quantities imported and exported are given in the form of equation (V.7), in which the percentage changes in composite quantity and price are the percentage change in total quantity demanded and weighed averages percentage changes in producer price, consumer price, import price and export price by using their cost share in the last year as weights, respectively. The elasticity of substitution σ = 2.2 (FAO).8

itimportitit

importit ppX τσ ˆ)ˆˆ(ˆ +−=

itort

ititort

it ppX τσ ˆ)ˆˆ(ˆexpexp −−−=

Xnetimportit = Ximportit – Xexportit pimportit = pibit(1 + PSEimportit) pexportit = pxbit(1 + PSEexportit) pibit = XRtpcifit pxbit = XRtpfobit Where: Ximport: Import. Xexport: Export. Xnetimport: Net import. XR: Exchange rate. prural: Rural consumer price. pcif: CIF price. pfob: FOB price PSE: Producer subsidy expenditure. i: index individual grain, including: rice, wheat, maize, sweet potato, potato, other coarse

grain, soybean, sugar, pork, beef, mutton, poultry, egg, and fish.

5. Market Clearing

Total supply equals total demand.

8 CIF price and FOB price are subject to changes in the prices of traded commodities in world market which are estimated based on projection of world market prices made by World Bank. Exchange rate is given in Appendix 1 Table 1.

Xnetimportit + SGit = DGit +Bit – Bi(t-1) 3. The formation of the equilibrium among endogenous determinants in CAPSiM model

system

CAPSiM is driven by endogenous determinants of supply and demand. Supply equations, which are decomposed by area and yield, grain and meat, allow producer

own price and cross market responses as well as the effects of shifts in technology stock of agriculture, irrigation stock, ratio of erosion area to total land area, ratio of salinity area to cultivate area, yield change due to exogenous shock of climate, and yield change due to other exogenous shock.

Demand equations, which are decomposed by urban and rural areas, grain and meat, allow consumer own price and cross market responses as well as the effects of shifts in income, population level, market development and other shocks. The total change in supply of different commodities between periods is derived by first order derivative of equation 1.1 (for grain) or equation 1.2 (for meat). It can be stated in terms of its components, as follows:

Salinityt

Salinityt

iErosiont

Erosiont

it

ti

t

ti

tj

jtsj

ij-ti

itsi

-ti

it

ZZ

hZZ

gI

Ik

RR

cP

P

PP =

SS

1111)1()1()1(

)1(−−−−−≠

Δ+

Δ+

Δ+

Δ+

Δ+

ΔΔ ∑ δδδδηη

where: Sit = market supply curve for ith commodity; ηsi = own price elasticity of supply of ith commodity; ηsj = cross price elasticity of supply of ith commodity; δ = 1 for grain and 0 for meat; In a similar fashion, total change in demand for different commodity between periods is derived by first order derivative of equation 2.1 or equation 2.2

MKTt

MKTt

t

t

ij tj

jtdj

ti

itdi

ti

it

ZZ

im + YY

ie + PP

+ P

P =

DD

11)1()1()1(

)2(−−≠ −−−

ΔΔΔΔΔ ∑ δξξ

where: Dit = domestic demand curve for ith commodity; ξdi = own price elasticity of demand for ith commodity; ξdj = cross price elasticity of demand for ith commodity; δ = 1 for rural, 0 for urban. Figure V.6 below graphically illustrates price determination in period t+1. Finding the equilibrium values, P1* and Q1* requires knowing S1 and D1 first. Recall from (1) and (2) above that for the ith commodity, when ΔPit = 0, the following is evident:

Salinityt

Salinityt

iErosiont

Erosiont

it

ti

t

ti

tj

jtsj

ijt

Pt

ZZ

hZZ

gI

Ik

RR

cP

P =

QS

it

1111*

)1(*

1

0|)3(−−−−−≠−

=Δ Δ+Δ

+Δ

+Δ

+ΔΔ

∑ δδδδη

MKTt

MKTt

t

t

ij tj

jtdj

t

Pt

ZZ

im + YY

ie + PP

= Q

D it

11*

)1(*

1

0|)4(−−≠ −−

=Δ ΔΔΔ

ΔΔ∑ δξ

With such relationships, Figure V.6 shows S1 and D1 can be deduced as follows:

]1[)5(0

1

0

1*0

1

0

1*0

1*01 Salinity

Salinity

iErosion

Erosion

iiiij j

jsj Z

ZhZZg

IIk

RRc

PP

Q S Δ+Δ+Δ+Δ+Δ

+= ∑≠

δδδδη

]1[)6(0

1

0

1*0

1*01 MKT

MKT

iiij j

jdj Z

ZmYYe

PP

Q D Δ+Δ+Δ

+= ∑≠

δξ

Figure V.6 Note that (5) and (6) simply add Qo* and the changes in demand or supply attributed to shiftors, i.e. variables other than own price which is held constant at Po*. Once S1 and D1 are known, the equilibrium values are solved using the following equations:

P+P=P that such)D-S()S-D(P=P 1*o*1d1s1

11*o

1 ΔΔ ξη)7(

]PP+[1S = ]

PP+[1D = Q *

o

1s1*

o

1d1

*1

ΔΔηξ)8(

From hereon, the cycle of calculations repeats in a recursive, dynamic fashion for the entire simulation period.

Price St St+1

P*1 P*0

Dt+1 Dt Quantity Q*0 S1Q*1D1

Demand Elasticity

Almost Idea Demand System (AIDS) (Deaton and Muellbauer，1980) is used as theoretical framework of demand research in CAPSiM. The general form of AIDS can be expressed as follows:

wi = Ai + Bilog(X/P) + Σjrijlog(pj) (1) Where, i, j = 1, …, n; wi is the cost share of commodity i in total expenditure, X. pj is

commodity j’s price, P is price index, defined as: log(P) = a0 + ΣkAklog(pk) + 1/2ΣkΣjrkjlog(pk)log(pj) (2) a0, Ai, Bi and rij are parameters to be estimated. Because the mode of food consumption is

subject to the influence of non-income and non-price variables, i.e., structure factors, the change (Z) in the structure factors should be taken into account in AIDS. Hereon, we assume that the parameters Ai and Bi in equation (1) and (2) are linearly related to zj，j = 1, …, n：

Ai = ai + Σsaiszs (3) Bi = bi + Σsbiszs (4) Where Z = (z1, …, zm); ai, ais, bi and bis are parameters to be estimated. According to properties of demand theory, we have the following constraints conditions for

equation (1) ∼ (4): adding up: Σiai = 1 (5a) Σiais = Σibis = 0，s = 1, …, m (5b) Σibi = 0 (5c) Σirij = 0 (5d) Homogeneity: Σjrij = 0 (6) Symmetry: rij = rji，i ≠ j (7) The stochastic structure forms of equation (1)∼(4) can be obtained by adding subscript h

representing household’s characteristics, and adding an error term εih in equation (1). The demand system above is of nonlinear form. Suppose that the distribution of εih is

multivariate normal. The parameters in above nonlinear equation system can be estimated by ML estimation of sufficient statistics. In this estimation, homogeneity and symmetry are imposed.

There are 8 components in Z: 3 dummy variables (town, city and occupation) and 5 demographic variables (family size, age groups, etc.). Because the dependent variable is budget (cost) share, covariance matrix is singular and one of these equations must be cancelled.

Consumer’s income elasticity (eiy), uncompensated price elasticity (eij) and compensated price elasticity (ceij) can be expressed as:

eiy = 1 + (bi + Σsbiszs)/wi (8) eij = -δij + rij/wi – (bi + Σsbiszs)[ ai + Σsaiszs +Σkrkjlog(pk)]/ wi (9) ceij = eij + wjeiy (10)

δij is Kronecker delta. In equation (8)∼(10), the expenditure of demand and price elasticity vary when

urbanization level, occupation, family structure, and other factors change. Mathematically, the effect of zi can be expressed as: Keeping other variables unchanged, to derivate equation (1) with respect to the mth zi to

get: dwi = aimdzm – (bi + Σs≠mbiszs)[ Σjajslog(pj)]dzm (11)

- {bimΣjajslog(pj) +bim[logx – a0 – Σj(aj + Σs≠majszs)log(pj)

-ΣiΣjrijlog(pi)log(pj)]dzm} The changes in income and price are embodied in equation (11). Dividing (10) with wi, we

get the percentage change in consumption of the ith commodity (qi) at different zm level, i.e.: dwi/wi = dqi/qi (12)

= {aimdzm – (bi + Σs≠mbiszs)[ Σjajslog(pj)]dzm - {bimΣjajslog(pj) +bim[logx – a0 – Σj(aj + Σs≠majszs)log(pj)

-ΣiΣjrijlog(pi)log(pj)]dzm}}/wi The price ealticities for urban and rural households used in CAPSiM are Marshall uncompensated since we are more interested in substitute effects among commodities when comparative prices change. The following two tables give these elasticities estimated from household data in 2005.

Table 2-1 Urban household’s Marshall uncompensated price elasticities

Milled rice

Wheat Maize Sweet -potato

Potato Other Coarse

Soybean Cotton Edible Oil

Sugar Vegetables Fruits

Milled Rice -0.22579 0.03991 0.00014 0.00017 -0.00027 -0.00046 -0.00153 0.00669 0.00757 -0.00672 -0.06062 0.00664 Wheat 0.06015 -0.33352 0.00015 0.00015 -0.00103 -0.00049 -0.00303 0.01817 0.00986 -0.00413 -0.05373 -0.00109 Maize 0.00938 0.00629 -0.27686 0.00515 0.00094 0.01941 0.00028 0.02402 0.01144 0.00011 -0.05285 0.01048 Sweet-potato 0.00960 0.00594 0.00443 -0.43984 -0.00227 0.01414 0.00033 0.01640 0.00253 -0.00172 -0.04248 -0.00597 Potato -0.00767 -0.01030 0.00004 -0.00085 -0.16484 0.00396 0.00357 0.02700 0.00649 -0.00082 -0.16942 0.02687 Other Coarse -0.00342 -0.00256 0.00810 0.00687 0.00789 -0.09103 0.00005 0.03043 0.01450 0.00032 -0.06531 0.01375 Soybean -0.01284 -0.01367 -0.00015 -0.00018 0.00179 -0.00053 -0.39719 0.00635 0.00500 0.00125 -0.00535 0.01038 Cotton 0.00364 0.01498 0.00067 0.00043 0.00407 0.00209 0.00155 -0.37648 0.02828 0.01025 -0.00207 0.01030 Edible Oil 0.00793 0.00784 0.00026 -0.00011 0.00107 0.00082 0.00164 0.03060 -0.74139 -0.01858 -0.04127 -0.00376 Sugar -0.04942 -0.02173 -0.00020 -0.00055 -0.00063 -0.00050 0.00131 0.04069 -0.07111 -0.52302 -0.00144 0.03352 Vegetables -0.02357 -0.01421 -0.00056 -0.00057 -0.00509 -0.00166 -0.00026 0.00129 -0.00741 0.00017 -0.44822 0.01664 Fruits -0.00232 -0.00432 -0.00008 -0.00034 0.00132 -0.00019 0.00090 0.00428 -0.00178 0.00286 0.02473 -0.59471 Pork -0.01431 -0.00902 -0.00020 -0.00029 -0.00006 -0.00040 0.00073 -0.00039 0.00106 0.00076 0.01332 -0.02791 Beef -0.00583 -0.00586 -0.00007 -0.00020 -0.00155 -0.00002 0.00132 0.00451 0.00928 0.00118 -0.01054 -0.02793 Mutton -0.00521 -0.00639 0.00004 -0.00012 -0.00260 0.00035 0.00259 0.00899 0.01796 0.00234 -0.00554 -0.02722 Poultry -0.01272 -0.00899 -0.00007 -0.00021 -0.00054 -0.00015 0.00067 -0.00251 0.01363 0.00146 -0.03149 0.00567 Eggs -0.00588 -0.00737 0.00005 -0.00018 -0.00039 0.00023 0.00180 -0.00158 0.00966 0.00193 -0.00408 -0.01126 Milk 0.00527 0.00266 0.00013 0.00004 0.00052 0.00040 0.00121 -0.00495 -0.00947 0.00085 0.02363 -0.01762 Fish -0.00291 -0.00281 -0.00008 -0.00014 0.00027 -0.00014 0.00256 0.00326 0.00353 0.00121 0.00499 -0.00064 Other Food -0.01451 -0.01055 -0.00041 -0.00048 -0.00069 -0.00104 -0.00172 -0.00374 -0.00652 -0.00145 -0.04144 -0.02680 Non-food -0.00976 -0.00519 -0.00022 -0.00013 -0.00118 -0.00069 -0.00226 -0.00788 -0.00261 -0.00129 -0.02592 -0.00742

Table 2-1 Urban household’s Marshall uncompensated price elasticities (continued) Pork Beef Mutton Poultry Eggs Milk Fish Other food Non-food Milled Rice -0.03723 0.00239 0.00076 -0.00707 -0.00114 0.02226 0.00958 0.01364 0.33109 Wheat -0.03397 -0.00017 -0.00138 -0.00817 -0.00742 0.02046 0.00675 0.00113 0.43131 Maize 0.00270 0.00733 0.00569 0.00998 0.01215 0.02721 0.01528 0.01274 0.49915 Sweet-potato -0.00601 0.00397 0.00270 0.00373 0.00202 0.02193 0.01183 0.01254 0.76122 Potato 0.00266 -0.01047 -0.01211 -0.00408 -0.00483 0.01543 0.01113 0.03604 -0.04782 Other Coarse 0.01736 0.01125 0.00874 0.01269 0.01690 0.03190 0.02080 0.01909 0.44166 Soybean 0.02282 0.00899 0.00738 0.00789 0.01121 0.01668 0.03354 0.02003 -0.02339 Cotton -0.00859 0.00741 0.00689 -0.00550 -0.00610 -0.00718 0.01148 0.02216 -0.21831 Edible Oil 0.01389 0.01661 0.01569 0.03403 0.01972 -0.01488 0.01807 0.00021 0.35163 Sugar 0.02235 0.00858 0.00760 0.01410 0.01341 0.01381 0.01929 0.00465 0.08929 Vegetables 0.02607 0.00061 0.00021 -0.00939 -0.00116 0.01870 0.00925 -0.01391 0.17807 Fruits -0.06804 -0.01061 -0.00674 0.00605 -0.00853 -0.00705 0.00242 -0.03983 0.32197 Pork -0.48279 0.01824 0.01407 0.01072 -0.00545 0.01255 0.03223 -0.01490 0.05202 Beef 0.08647 -0.82323 0.01234 0.08648 0.01004 0.00500 -0.00163 0.01795 0.04227 Mutton 0.12985 0.02329 -0.85573 0.08451 0.00025 0.00684 -0.00394 -0.00105 0.18080 Poultry 0.03124 0.05659 0.03054 -0.70081 0.06002 0.00731 0.04596 0.00436 0.00004 Eggs -0.01506 0.01028 0.00100 0.06960 -0.59360 0.04421 0.02791 -0.01252 0.18526 Milk 0.02088 0.00144 0.00055 0.00278 0.03095 -0.92967 -0.01324 -0.02181 0.15543 Fish 0.06338 0.00040 -0.00120 0.02861 0.01276 -0.00441 -0.74215 -0.01805 0.15156 Other Food -0.05265 -0.00461 -0.00454 -0.01082 -0.01502 -0.01126 -0.02270 -0.42987 -0.53920 Non-food -0.04113 -0.00574 -0.00286 -0.01166 -0.00920 -0.00379 -0.01247 -0.09524 -0.96496

Table 2-2 Rural household’s Marshall uncompensated price elasticities

Milled rice

Wheat Maize Sweet -potato

Potato Other Coarse

Soybean Cotton Edible Oil

Sugar Vegetables Fruits

Milled Rice -0.37040 0.00155 0.00158 -0.00037 0.00115 -0.00180 0.00071 0.01419 0.00265 -0.00079 -0.04031 -0.00069 Wheat -0.00007 -0.34222 0.00184 -0.00033 0.00359 -0.00136 -0.00046 0.02481 0.00775 -0.00096 -0.05152 -0.00062 Maize 0.03802 0.02706 -0.30506 0.00408 0.00769 0.04323 -0.00086 0.03536 0.00988 -0.00061 -0.05643 0.00480 Sweet-potato -0.01158 -0.00377 0.03515 -0.07963 0.01059 0.05790 -0.00072 0.04013 0.00166 -0.00298 -0.07287 -0.00945 Potato 0.04399 0.09433 0.02745 0.00449 -0.26792 0.00399 0.00475 0.02877 0.03328 -0.00098 -0.11204 0.01799 Other Coarse -0.00352 0.00384 0.06377 0.00994 0.00223 -0.16996 -0.00179 0.04189 0.01181 -0.00054 -0.06361 0.00745 Soybean 0.01473 -0.01491 -0.00551 -0.00062 0.00376 -0.00609 -0.51400 0.03482 0.00465 0.00050 -0.01235 0.00558 Cotton 0.00580 0.02224 0.00359 0.00052 0.00113 0.00318 0.00197 -0.79466 0.00393 0.00066 0.00612 0.01018 Edible Oil -0.00409 0.00952 -0.00007 -0.00032 0.00238 0.00003 0.00036 0.01199 -0.62862 -0.02616 -0.06233 0.00241 Sugar -0.04212 -0.02785 -0.00517 -0.00139 -0.00100 -0.00370 0.00022 0.01276 -0.24439 -0.47464 0.00126 0.03737 Vegetables -0.03980 -0.02764 -0.00637 -0.00088 -0.00188 -0.00495 -0.00033 0.00798 -0.01245 0.00015 -0.42118 0.01643 Fruits -0.02576 -0.01426 -0.00225 -0.00087 0.00121 -0.00122 0.00025 0.01831 -0.00028 0.00457 0.08590 -0.75157 Pork -0.03381 -0.02098 -0.00576 -0.00073 -0.00027 -0.00317 -0.00033 0.00117 0.00214 0.00030 -0.04031 -0.01857 Beef -0.01614 -0.01545 -0.00365 -0.00045 -0.00957 -0.00210 0.00049 0.02031 0.04155 0.00174 -0.04464 -0.01796 Mutton 0.00244 -0.00209 -0.00121 -0.00007 -0.00149 -0.00022 0.00123 0.02039 0.04108 0.00202 -0.03989 -0.02617 Poultry -0.01688 -0.01811 -0.00434 -0.00042 -0.00051 -0.00309 0.00041 -0.00411 0.04723 0.00164 -0.04394 0.02914 Eggs -0.00743 -0.01581 -0.00234 -0.00035 -0.00030 -0.00150 0.00088 0.00001 0.01525 0.00165 -0.06226 -0.00331 Milk -0.00588 -0.00230 0.00108 0.00001 -0.00011 0.00043 0.00047 -0.04229 -0.09177 0.00053 -0.00949 -0.01271 Fish -0.00573 -0.01335 -0.00177 -0.00010 -0.00001 -0.00073 0.00120 0.00665 0.01003 0.00125 -0.05040 0.01127 Other Food -0.03500 -0.01942 -0.00475 -0.00044 -0.00012 -0.00344 -0.00054 0.00563 0.00077 -0.00039 -0.01560 -0.00375 Non-food -0.04167 -0.02773 -0.00394 -0.00061 -0.00161 -0.00320 -0.00132 -0.01331 -0.01092 -0.00103 -0.06179 -0.00613

Table 2-2 Rural household’s Marshall uncompensated price elasticities (continued) Pork Beef Mutton Poultry Eggs Milk Fish Other food Non-food Milled Rice -0.01611 0.00115 0.00380 0.00174 0.00254 0.00641 0.01302 0.00100 0.27898 Wheat -0.01978 -0.00021 0.00267 -0.00241 -0.00402 0.00636 0.00575 0.00241 0.21877 Maize -0.03031 0.00070 0.00443 -0.00125 0.00319 0.01153 0.01567 0.00430 0.48458 Sweet-potato -0.03441 0.00071 0.00567 0.00191 0.00164 0.01033 0.02080 0.02004 0.35890 Potato 0.00973 -0.04731 -0.00590 -0.00104 -0.00090 0.00591 0.01308 0.02916 -0.08084 Other Coarse -0.00416 0.00210 0.00646 -0.00040 0.00531 0.01155 0.02081 0.00647 0.45040 Soybean 0.00169 0.00395 0.00791 0.00760 0.01151 0.00800 0.02794 0.00768 0.11315 Cotton -0.00702 0.00573 0.00642 -0.00513 -0.00491 -0.01147 0.00987 0.02221 0.01964 Edible Oil 0.02594 0.01816 0.02005 0.04653 0.01687 -0.03491 0.02334 0.02898 0.19993 Sugar 0.02153 0.00739 0.00940 0.01554 0.01527 0.00696 0.02339 0.00680 0.19236 Vegetables -0.02509 -0.00205 -0.00188 -0.00534 -0.01227 0.00450 -0.00310 0.01340 0.14777 Fruits -0.10810 -0.00815 -0.01340 0.03173 -0.00656 -0.00248 0.02353 -0.00194 0.22134 Pork -0.50216 0.00773 0.01162 0.01851 0.01136 0.00510 0.02663 0.00520 -0.01368 Beef 0.09170 -0.74365 0.10223 0.09357 0.04078 -0.00036 -0.00977 0.01996 -0.14858 Mutton 0.12492 0.09163 -0.85695 0.06280 0.01889 0.00055 -0.00738 0.00425 -0.03472 Poultry 0.09893 0.04077 0.03065 -0.78608 0.04166 0.00066 0.03894 0.00460 -0.00717 Eggs 0.06487 0.01673 0.00939 0.03874 -0.74832 0.05448 0.02210 0.01902 0.19851 Milk 0.01108 -0.00375 -0.00326 -0.00724 0.11802 -0.95956 -0.00947 0.00764 -0.14143 Fish 0.07732 -0.00453 -0.00425 0.02314 0.01037 -0.00112 -0.78703 0.00342 -0.12563 Other Food -0.01233 0.00090 -0.00088 -0.00246 -0.00118 0.00330 0.00222 -0.60579 -0.10675 Non-food -0.06746 -0.00662 -0.00600 -0.01246 -0.01072 -0.00344 -0.01579 -0.04587 -1.06160

An Analysis on the Effect of the Distribution Control System for Tangerine

Seong Bo KOProfessor of Applied Industrial

Economics at Juju National University

Ⅰ. IntroductionCitrus Unshiu (the satsuma, satsuma tangerine, mikan, Jeju satsuma) producers,

producer cooperatives and distributors (on the Jeju Island) voluntarily agreed to adopt the "mandate for controlled satsuma distribution" based on the "act on agricultural product distribution and price stabilization" in October 2003 for field-grown satsumas in response to a drastic drop in the fruit's prices for four consecutive years from 1999 to 2002.

The mandate was spreaded nationwide since 2004 with government registered wholesalers/wholesale markets continuing to adopt the mandate to those satsumas produced until 2006. The results were field-grown satsuma price stabilization for four years between 19999 and 2002 and increased crude income for three consecutive years from 2004 to 2006 to over KRW 600 billion. And it seemed as if the "formula (the mandate = significant increase in crude income from satsuma = regional economic promotion)" could be set up.

Although the mandate was adopted with the same contents as that of 2006 to those field-grown satsumas produced in 2007 and 2009, the years of overproduction, the price of 10kg of the satsuma dropped an average of 20% (for those produced in 2009) and 40% (for those produced in 2007) compared to the average price between 2003 and 2006. Consequently, the above formula (the mandate = price hike) was shown to be not effective, which demonstrated that the mandate was no longer the "panacea" for price increase. Therefore, this formula valid from 2003 to 2006 needed to be analyzed quantitatively in detail in comparison to the years 2007 and 2009 when satsuma prices plummeted.

The purpose of this study was (1) to analyze the background/causes behind the introduction of the mandate for the first time in 2003 and (2) to quantitatively analyze the effect of mandate adoption on satsuma prices and farm incomes.

Ⅱ. Significance and background on the introduction of the mandate for controlled satsuma distribution

1. Significance and effect of the mandate

The mandate for controlled satsuma distribution was a voluntary self-help distribution

tool/program by producer cooperatives to legally control producers and middle distributors with the approval by the Minister for Food, Agriculture, Forestry and Fisheries in order to improve farm incomes and distribution efficiency by controlling the amount or quality of products distributed in markets.

Generally, the following are the benefits and drawbacks consequent to the adoption of distribution mandates but it is not easy to quantify the actual effect of this mandate program and whether it was carried out effectively.

The benefits of the distribution mandate include ① market and price stabilization by reducing market uncertainties by supporting markets, ② farm income and satsuma price increases in a stable basis, ③ an orderly distribution of agricultural products, ④ transfer of market power from distributors to producers, ⑤ increased information in markets and increased market efficiency, ⑥ quality improvement by stimulating farmers to produce quality agricultural products and guarantee quality products to consumers, and ⑦ increased demand from R&D on markets, product development and marketing.

The drawbacks include ① increased consumer prices, ② decreased choice for consumers, ③ the issue of free riders, ④ significant administrative, supervisory and monitoring costs, ⑤ difficulty in resolving differences in interest among farms and regions, and ⑥ stagnation in improving quality and product variety by taking the program for granted. Especially, the issue of free riders props up because it is difficult to completely control the distribution amount so that it is not an exaggeration to say that the success of this mandate program heavily depends on minimizing free riders because they can reduce the effect of mandate or completely remove the effect in the worst case.

2. Background for the introduction of the mandate for controlled satsuma distribution (for satsumas produced in 2003) for the first time

A. Causes of satsuma price plummeting and decline in the Jeju economy

The Jeju economy suffered as prices of field-grown satsuma produced for four consecutive years from 1999 to 2002 plummeted as never seen before.

Firs, this continued decline decreased the crude income from satsuma by 40% from KRW 515.8 billion in 1998 to KRW 316.4 billion in 2002. Second, the average farm income also decreased from KRW 39 million in 1996 to KRW 22~24 million since 1999 to 2001 but increased slightly because of vegetable price increases to KRW 293 million in 2002. Third, the average debt per each farm on the Jeju Island, which remained the national average at less than KRW 13 million until 1997, started to increase drastically since the financial crisis of Korea to KRW 30.84 million in 2001 and to KRW 32.53 million in 2002, which was a 36% increase compared to the national average of KRW 24 million. Fourth, this sharp decline in satsuma prices drastically decreased the GRDP ratio of the primary industry of the Jeju Island from 25.3% to 16.7% during the same period, and the per capita GRDP to KRW 8.85

million, which was only 78.4% of the national average at KRW 11.28 million. The problem is compounded from the fact that the ratio of workers in the primary industry on the Island was high at around 25%.

Then, what was the cause of plummeted satsuma prices? Although satsuma farmers point out the culprit to be over supply due to imported oranges and satsuma over production, exports indicated poor quality of satsuma and over supply of fruits other than this tangerine. On the other hand, along with these causes, wholesalers in Garak whole sale market (the largest fruit and produce wholesale market in Korea) pointed out a decline in satsuma flavor, consumption decline, production of fruits alternative to the satsuma such as strawberries, and poor quality satsuma due to such factors as rottenness, as some of the causes of satsuma price decline. In other words, there was a gap between producers and consumers for the causes of price decline with the former pointing out the amount of production as the basic problem of the satsuma industry, whereas the latter, the quality. Consumers thought the basic problem lied in quality and the amount of production thought to be important for producers was not such a serious problem to the point of being behind distribution.

In other words, the satsuma suffered two-tier problems in which it had less quality competitiveness compared to other fruits (domestic and imported oranges) and decreasing cost competitiveness in which decreasing prices of other fruits brought about price decreases in other fruits as well as the satsuma. The sugar-acid ratio used for the quality of field-grown satsuma decreased significantly since 1999 to less than 9.0 compared to 1998 (the year when the crude income from the satsuma was over KRW 500 billion) at 9.8 and 1996 at 10.0.

Comparison of the quality and crude income per year from field-grown Onju mikan grown on the Jeju Island

Year Sweetness (。Bx) Sourness (%) Sugar-acid ratioProduction amount

(ton)

Crude income (in KRW 100

million)

’94 10.8 1.2 9.0 548,945 5,521

’95 9.9 1.4 7.1 614,770 4,334

’96 11.1 1.1 10.0 479,980 6,079

’97 11.2 1.2 9.3 693,200 4,009

’98 9.8 1.0 9.8 543,980 5,158

’99 8.9 1.0 8.9 638,740 3,257

2000 9.8 1.1 8.9 563,341 3,708

2001 10.3 1.16 8.9 646,023 3,617

2002 9.6 1.36 7.1 788,679 3,164

평균 10.15 1.16 8.8 613,073 4,316Source : Report from the satsuma testing field

Plunge in satsuma prices and decline in Jeju regional economy Unit : in KRW 10,000, %

Year Farm income Farm debt

Regional gross domestic product (GRDP) related

Ratio of the primary industry

Per GRDPComparison to

national average (=100)

1993 2,113 694 27.2 521.2 85.0

1994 2,217 809 27.3 626.4 90.1

1995 2,722 855 31.0 720.4 90.8

1996 2,990 1,163 24.8 766.5 87.6

1997 2,238 1,298 24.1 845.3 90.0

1998 2,596 1,869 25.3 780.6 85.3

1999 2,224 2,572 25.7 860.0 85.7

2000 2,473 2,934 22.4 881.2 82.4

2001 2,476 3,084 16.7 885.5 78.4

2002 2,930 3,253 - - -

Sources : "Report on farm economy statistics" from MIFAFF and "Report on GRDP" by Statistics Korea for each year.

B. Limitations and problems of ordinance related with the production and distribution of Jeju Island satsuma

Until now programs related with Jeju satsuma were adopted upon the "ordinance related with the production and distribution of Jeju Island satsuma", which was based on the "special law to transform Jeju Island into an international free economic zone" (article 50 on stable supply and demand on agricultural, forestry and fishery products). However, the ordinance is carried out not in tune with the "act on distribution and price stabilization of agricultural, forestry and fishery products" of the central government (MIFAFF), central to stable supply and demand of these products. In other words, the act allows the enforcement and policy support for all satsuma-related organizations including sellers and satsuma farms. On the other hand, at the legal interpretation level, the ordinance can be used to punish those who illegally distribute non-standard Jeju satsuma in Garak wholesale market; however, it is a regional law effective in limited terms because it is based on the regional special law of Jeju Island.

Comparison of benefits and drawbacks of satsuma policy implementation based on the ordinance and the act

Category Local government-oriented satsuma policies based on the ordinanceSatsuma policies at national level based on

the act

Applicable period Temporary (2011) No limitation

Applicable region Jeju province Nationwide

Policy consistency Lack of consistency, high potential Lack of consistency, low potential

Enforcement commitment and

effectivenessRelatively low Relatively low

Usability in mobilizing policy

support funds

Not easy due to disagreement between policy enforcement platform and support

funds (agriculture stabilization funds)

Easy due to agreement between policy enforcement platform and support funds

(agriculture stabilization funds)

Safety and practicability Low (amendable)

High(can differ each year based on mandate order)

Policy direction Top down Bottom up(Self-funding, mandate, etc)

Perception by Jeju people

Recognized as an autonomous ordinance not having to be abided by Recognized as enforceable legal policies

It is true that the ordinance was amended due to many reasons such as changes in local government heads as Korea transformed from a centrally oriented government to one that granted more power to local governments. And this change was believed to be an important factor in the ordinance being operated at a short-term basis rather than medium to long-term basis and lack of consistency in policies based on the ordinance. And of course, there are free riding farmers and merchants who tried to use the loopholes innate to the ordinance such as the currently loose legal system and lack of strong committment to enforce the law related with coming elections to choose local government representatives.

This reali